Availability, Reliability and Maintainability
Ian D. MacKenzie Process & Energy Business Unit Peter Brotherhood Ltd, Peterborough England, U.K.
The Recip – A State of the Art Compressor
Abstract:
The reputation of the Reciprocating Compressor has suffered over the years due to prejudices and misconceptions regarding various aspects of its design and construction. Furthermore, there is a belief amongst operators, rotating equipment engineers and buyers within the industry, that they are generally unreliable and are at best difficult and expensive to maintain. Prejudices have been perpetuated by amongst others, manufactures of alternative designs of compressor. In the past, much of this bad reputation may well have been deserved, however, in my opinion this is no longer the case. In this paper I discuss these prejudices and try to separate some of the facts from fiction.
1 Introduction The *A.R.M. Criteria is a relatively new concept, being introduced some ten years ago. However, its basic principles have been relevant since the first compressors can into existence.Although the three criteria are applied together and are inter-related, each has a quite specific meaning which is sometimes misunderstood. In this paper I will first give the most commonly accepted definitions for A, R and M, then go on to discuss the factors that can affect them. Prejudices and misconceptions clearly exist with regard to the Recip, especially when being compared to the alternative designs.These can vary from minor differences of opinion which have no real influence on the machine type selection, to very significant difference of opinions which may result in a Recip being effectively precluded from consideration.For example, when for a critical service, 2 Recips are specified as an alternative to 1 Screw or 1 Centrifugal compressor. There is no doubt that in the past the Recip has deserved some of it’s bad reputation for a variety of reasons which I will discuss later in this paper. Indeed, in the early 1970’s, most compressor manufacturers, including our company, were experiencing all kinds of problems, some of which resulted in Recips being replaced by alternative designs.The Recip’s reputation has probably suffered worst in the Offshore Oil and Gas industry in the North Sea, where any problems are greatly exacerbated by the shear logistics of problem solving.Probably the most extreme and damaging case history for the Recips reputation is that of the ill-fated Piper Alpha. The Recips were initially implicated as a possible cause of the disaster, as it had been documented that they had suffered from several problems. They were later, totally cleared of having made any contribution; however, the compression modules on the replacement platform, Piper Bravo, were equipped with Centrifugal Compressor trains. In this paper, I will discuss some of the problems that our company has experienced and how we have dealt with them and hopefully try to restore some confidence in the muchmaligned Recip. It should be noted that I have only considered the highest specified compressors, typically those in accordance with API 618. * Throughout the paper I will in many cases, abbreviate Availability, Reliability and Maintainability and refer simply to A, R & M. 2.1 Availability & Reliability Generally speaking the first two criteria (A & R) are considered together and the third, Maintainability, although equally important, is usually separated from the other two. The reason for this is that with A & R the definitions can be clearly specified and in many cases now can be included in guarantee clauses within a contract whereas with M only the principles and guidelines can be specified, as there can be many variables and constraints affecting the designer’s ability to meet the ideal solution in all cases. Although there are some variations of the definitions of A & R, the most commonly accepted definitions are as follows: Availability ( % ) = 8760 – SD – UD . 100 8760 Where
The Essay on Case Problem: Stateline Shipping and Transport Company
Rachel is considering using each of the plant and waste disposal sites as intermediate shipping points. Now, there is a possibility to reduce the transportation coat by utilizing alternative routes. For example, as per the given cost table, cost to transport from Kingsport to Duras is $17. Instead of this route, if we transport first from Kingsport to Macon and then from Macon to Duras the cost is ...
The Term Paper on Analysis and Design of Software Architecture
Outline 1 2 3 4 5 6 7 8 Development Process Requirements Quality Attributes Runtime QA Non-runtime QA Requirements Analysis: Example Architectural Analysis & Design Architectural Views Denis Helic (KMI, TU Graz) SA Analysis and Design Oct 19, 2011 2 / 78 Development Process Methodology Different software development processes have software architecture as a part of the process Rational unified ...
SD – Scheduled Downtime UD – Unscheduled Downtime 8760 – the Hours in 1 Year Similarly : Reliability ( % ) = 8760 – SD – UD . 100 8760 – SD The above of course assumes continuous running. Should the running period be only say 4000 Hours by design, then this figure should be substituted in the formulae. The subtle difference between these formulae means that even if there is no Unscheduled Downtime, the Availability cannot be 100 %.There is a school of thought that only the Unscheduled Downtime should be considered and therefore 100 % can be achieved. In other words, the machine is Available for 100 % of the time that it was planned to be. This means, however, that there is no difference between the definitions of A & R, which of course makes no sense. 2.2 Maintainability Maintainability, the ability to maintain a machine is really self-explanatory; however, it does have a direct affect on Availability. Quite simply, if the Scheduled Downtime is extended because the machine is difficult to maintain, Availability is compromised. 3 Factors Affecting A. & R. 3.1 Single or Spared Machines In some critical services, a minimum of 2 compressors are required, with 1 in operation and 1 on standby. This is usually dictated by the process licensor, or in some cases the engineering contractor, who may be required to give overall plant A & R guarantees. Provided the standby machine is well maintained and kept in a state of readiness, there is no reason why the Availability should not exceed 99 %. In fact, assuming no Unscheduled Downtime, then the Availability should be 100 %, which is the whole point of having spared machines. In some cases where an alternative design compressor, such as a Screw can handle the duty, 1 only may be specified, as an alternative to 2 Recips.This obviously puts the Recip at a disadvantage, when only capital costs are considered. However, the Screw alternative will attract a significant power penalty, which will go a long way in offsetting the cost of 2 machines.The question is however; can the Recip really be so unreliable compared to a Screw or a Centrifugal? I believe not and that decisions are based on perceptions and not necessarily the facts. 3.2 Basic Design Parameters 3.2.1 Gas & Net Rod Loads These are probably the most important parameters in determining whether a compressor will be reliable or not, as virtually all the major compressor components are affected by one or both of these parameters.
The Essay on The Postmodern Era of Graphic Design
The 1980s was characterized with the rise of eclecticism as an art trend. Style was no longer dictated by established rules but by the random convergence of different theories and practices. The aforementioned phenomenon, in turn, ushered the postmodern wave in graphic design. The liberal atmosphere that prevailed over their craft prompted many graphic designers to come up with more creative and ...
Furthermore, the consequences of exceeding these design parameters, or indeed selecting unrealistically high values when rating a compressor frame, can be disastrous. In some cases problems have occurred and subsequent investigations have failed to recognise that the primary cause has simply the fact that the compressor is not capable of withstanding the loads for which it has been apparently designed. This was particularly true in 1960’s and the early 1970’s when many so called “Fag Packet” designs were adopted with quite genuine belief in their accuracy, only to find the awful truth when standing next to the compressor in the middle of some refinery, surrounded by a group of irate customers. I might add that this practice was not confined to the Recip market)! Since then there has been a much more widespread use of verification techniques, such as Strain Gauge analysis and of course more recently the extensive use of Finite Element Analysis FEA. Since the introduction and world-wide acceptance of ISO 9001, which necessitates the requirement to fully document the verification of all major design calculations, the possibility of such a fundamental design fault being an issue, should have all but been eliminated. 3.2.2 Rotational Speed This is probably one of the most important factors affecting Reliability and more significantly, Availability. It is one parameter, however, that is sometimes completely ignored, being replaced as acceptance criteria by the universally used Mean Piston Speed (M. P. S.).However, I have deliberately separated these two. Although obviously related mathematically, they have two quite separate effects on A & R. (I would add that some companies, in particular the Oil Companies and Process Licensors, do actually specify limits for both).
Rotational Speed has the greatest effect on the residual Out of Balance Forces and Moments. Typically the compressor stroke is reduced in direct proportion to the speed, so that the M.P.S. remains within a given limit. The Inertia Force however, increases with the square of the speed, the formula for the Primary and Secondary Inertia Force being: F IN = m . ω 2 . R . ( Cos θ + R / L Cos 2θ ) where i.e. ω ω R L θ m – Mass of the reciprocating parts – Angular velocity in Radians / s = Speed . 2π / 60 – Crank throw ( Stroke / 2 ) in mm – Conn – Rod centres in mm – Crank angle in degrees
The Essay on Electrical Installation Design
E-power is a company that specializes in residential installation design at incredibly fair and competitive prices. We complete every electrical installation or design service safely, on time and on budget and with over 20 years’ experience in thus industry, we know exactly how to install systems that will save you money as well as help the environment. We provide a comprehensive range of design, ...
It can be seen that for a given reciprocating mass, if the Speed is doubled and the Stroke halved, the net effect is a doubling of the Inertia Force. In practice the reciprocating masses are usually less for higher speeds, however the net result is invariably higher Inertia Forces and their associated Moments. One other simple fact is sometimes overlooked. If the Speed of a given compressor is double that of another and their respective Mean Piston Speeds are the same, then the life expectancy of all the wearing components should be the same, with one notable exception, the Cylinder Gas Valves. Quite simply they have to open and close twice as many times in the higher speed machine.
Hence, if the valves have been designed using exactly the same design criteria, manufactured using the same processes and tolerances and from the same materials, then the Valve Plates will only last half as long. Modern Cylinder Gas Valves for use in the Refining and Petrochemical industries are typically designed such that the wearing parts give a minimum of 1 Year (8760 Hours) service before repair or replacement. [ See 3.3(B).1 ]. Normally this criterion is quoted by the manufacturers without giving any conditions regarding Rotational Speed. I would suggest that it is unreasonable to expect the Valves in a compressor running at say 1000 RPM to last as long as those in one running at say 300 RPM. In the long term, any components that do have a finite fatigue life, will of course have a service life in proportion to the inverse of the Speed. Hence the Availability cannot be as good for the higher speed compressor. 3.2.3 Mean Piston Speed ( M.P.S. ) The M.P.S. is limited mainly to ensure that the Piston and Bearer Rings and the Sealing Elements of the Piston Rod Packings can be expected to last the maximum possible time, typically 8760 Hours. Various figures are used in the industry; however, if no figure is stipulated in an enquiry, then we would use our own criteria of 4.6 M/s for Lubricated service and 3.6 M/s for Non – Lubricated service. We would expect the A & R would start to be compromised if these figures are exceeded. Generally speaking the lower the M.P.S. the better. 3.2.4 Cylinder Lubrication My comments in this section are applicable only to Horizontal Recips Apart from a few exceptions, compressors with Lubricated Cylinders will have higher values for A & R than those with Non – Lubricated Cylinders. The main reason is that with Lubrication a much wider choice of designs and materials are available to the designers of Cylinder Valves, Piston and Bearer Rings and Main Gas Packings. Also, the mechanism whereby these components work, is completely different on Non-lubricated Cylinders. See also 3.2 (B).
The Term Paper on Design and Analysis of a Laminated Composite Tube
A composite material is a material made from two or more constituents with significantly different physical or chemical properties. There are two main categories of constituent materials, matrix and reinforcement. At least one material of each type is required. (Composite Material). The phase distribution and geometry of the two materials have been ordered to optimise its properties, this has led ...
1, 2 & 3 ).
Many compressors these days are specified as Lubricated, but designed for future Nonlube service. In other words, the Lube machine is preferred, if the process can stand the carry-over of oil. However, they have an alternative that can be easily implemented, if this is later found not to be the case. Overall A & R can be affected by both too much or too little lubrication. Too much can result in problems with Valves, Piston Rings, Packings and in severe cases Pistons, due to a build-up of oil in the Cylinder. Too little lubrication can result in excessive wear of Piston and Bearer Rings, which in turn may cause the Piston to contact the Cylinder Liner, resulting in damage to both. Similarly Packing Elements can suffer from excessive wear as sections within the Packing may become totally starved of oil. In the case of metallic Elements, this would almost certainly result in Piston Rod damage. In the case of Cylinder Valves, with metallic Plates, they are usually guided by some nonmetallic device, during the operating cycle of the Valve. These rely on lubrication to work
successfully and should this be lost, then very rapid wear will commence. Some compressors are specified as requiring “Mini-Lubrication”. This is where the Lube rates are expected to be a fraction of those for full lubrication. I personally do not believe there is such a thing, although our company has supplied many machines to this specification. Essentially, the optimum lubrication rates have been established over the years by experimentation, but probably in the majority of cases, by trial and error. During commissioning, the lube rates are usually turned up to assist in flushing debris from the cylinders then turned down to a level where evidence of some oil starvation is evident. The rates are then increased to give a reasonable safety margin. Non-metallic Piston and Bearer Rings are designed for Lubricated service or Non-Lubricated service. The sliding mechanism of each type of Ring is fundamentally different. If a set of Rings that have been designed for Lubricated service, are starved of oil, they will begin to wear at a much greater rate than a set that has been designed to operate with no lubrication. They will also probably wear out faster than a set of metallic Rings in the same circumstances. To put a set of Rings into service, using lube rates that are simply based on a fraction of those for normal lubrication, is inviting problems. Good lubrication can be beneficial for several other reasons. It can act as a coolant in the Cylinder, so generally all the components will be operating at a lower bulk temperature than those on an identical service, without lubrication. It can also assist in flushing out wear products and foreign particles from the Cylinder, which may become trapped in one without lubrication. It can in some situations improve the sealing efficiency of Piston Rings, although this is still subject to debate. The type of lubrication system that is used can also have an effect on A & R, however, this is a subject that would require much more discussion than this paper would allow.Suffice to say that there are systems available that are very reliable and as such, any compromise of A & R can be avoided. 3.3(A) Major Component Design I have split this section into two parts. Part A deals with those components that are either considered to be strictly Non-Wearing components, or are subject to some wear, but their life expectancy would be 3 Years or better. Part B deals with those components that typically would not be considered to have a life expectancy greater than 3 Years. 3.3(A).1 Compressor Frame The design of the Crankcase, Distance Pieces and Extension Distance Pieces is obviously fundamental in determining whether a compressor will be Reliable or not. A badly designed Frame could at best give rise to excessive vibration, or in extreme cases result in catastrophic failure. In the late 1960’s and the 70’s, many problems were experienced, especially with HighSpeed compressors, with excessive machine vibration. Many of the usual excuses were
The Term Paper on How To Use CAD As A Design Tool
In order to use CAD effectively as a design tool, the designer must have the skills necessary to create, edit, and manipulate a 3D model of a part in order to create a realistic representation of an imagined object. In this activity you will build on the CAD skills that you learned in Activity 5.2b Introduction to CAD Skills. You will learn about and use additional tools and features available in ...
used such as “bad foundation design, inadequate structural steelwork, in the case of Offshore installations, it’s not operating on the design conditions”, etc., etc. In most cases however, the Frames were quite simply not designed to withstand the Dynamic loads to which they were being subjected. I suspect that as with our company, most calculations were carried out to determine the stresses and deflections of the Frame components, using Static Loads. In cases where Strain Gauge Testing was used to verify these calculations, the results often showed that the Frame was capable of even higher loads. The result was that, when subjected to the Gas Load exerted by the Cylinder through the Distance Pieces, the Crankcase side wall deflected. In most cases the side walls are asymmetrical due to the crank offset, so this deflection tended to be much greater on one side of the Distance Piece. This resulted in the cylinder moving from side to side. Once the relatively large mass of the cylinder, usually with pulsation dampener vessels attached, was set in motion, its inertia greatly exacerbated the problem. In many cases, the problem manifested itself in cracks appearing in the areas with the highest stress concentration, such as stud or bolt holes, in the Crankcase, but more commonly the Distance Pieces. Repairing or replacing these components and attempts to reduce the amount of Crankcase movement, would be expensive, but more importantly, extremely time consuming. Furthermore, this vibration could have a severe “knock-on” effect, with many other components having their useful life drastically reduced. Sadly, in what has always been a highly competitive market, most manufacturers did not have huge Research and Development budgets for Recip Compressors, so most bad experience was gained in the field, at the expense of the vendor, the contractor or most commonly the ultimate customer, the End User. In my opinion, this period had the most damaging effect on the reputation of the Recip. 3.3(A).2 Motion Work As with the Frame, if the Motion Work is badly designed then very little else matters. A Crankshaft has to be capable of withstanding a combination of Bending Stresses induced by the Rated Rod Loads and Torsional Stresses induced by the Transmitted Torque. Clearly if it has been designed incorrectly, or with insufficient safety margin, then the resulting failure could be catastrophic. Fortunately Crankshaft failures are few and far between, these days. However, due to the very nature of such a failure, where the costs and downtime can be horrendous, any occurrence does attract considerable attention. Conn-Rod assemblies by the nature of their design are subject to maintenance and as such do have an effect on A & R. The Conn-Rod Main Body and Cap are like the Crankshaft, expected to last the lifetime of the compressor. With the exception of the smallest frames, these are made from steel forgings, in many cases from drop, dye forgings. Over the years they have probably been subjected to more testing and experimentation than any other component. Hence failures, to the best of my knowledge, are extremely rare. However, most maintenance programmes include a periodic inspection of the Main and Crankpin Bearings. This entails removing the Conn-Rod cap, which in turn means loosening the Conn-Rod Bolts. In most compressors the Conn-Rod Bolts are a critical, if not the most critical component. They are subjected to the Net Rod Load, which invariably reverses from a maximum tensile load to a maximum compressive load every revolution. To minimise the fatigue
stress induced in the thread roots, these bolts are pre-tensioned with a load which exceeds the applied load. In many cases now, in particular on large bolt sizes, this pretension is carried out using Hydraulic Tightening techniques. As these are such a critical component they are normally inspected about every 3 Years running time, for signs of fatigue or corrosion and replaced if necessary. This does however, have only a very small effect on A &R. A Small End Bush is not such a critical component from a design point of view, however, it is much more susceptible to failure. This is mainly due to the fact that it depends on several other factors to operate successfully. Unlike the Main and Crankpin Bearings that have pressure forced oil feeds and operate on a Hydrodynamic oil film, the Small End Bush has to rely on Marginal lubrication. To avoid metal to metal contact, the oil has to flow into the clearance between the Bush and the Gudgeon Pin. For this to happen, the Rod Loading has to reverse from tensile to compressive once per crank revolution. During this period, the Gudgeon Pin migrates from on side of the Bush to the other and back again. Whilst there is clearance between Pin and Bush, the oil flows under pressure to cover the area that will then take the load when the Pin returns. This phenomenon is known as a “Rod Reversal” and is usually expressed in degrees of crank revolution, the maximum value being 180 Degrees. The design and operation of Small End Bushes is a highly complex subject and would require a complete paper to cover it, however, suffice to say that all the factors that may result in a failure need to be considered. A failure (seizure) of a Bush usually results in damage to the Crosshead Slippers, which may need re-metalling, damage to the Gudgeon pin, which may need replacing and in some cases the Small End Eye of the Conn-Rod may need opening out to accommodate an oversized Bush. Obviously if spares are not available, the downtime can be very lengthy. Crankpin and Main Bearings are usually kept as spares, mainly due to the fact that they can be easily damaged during maintenance or due to ingress of foreign particles into the oil system. However, we do not consider them as “Wearing Parts” and in many cases, where the compressors have been well maintained, these have lasted the lifetime of the compressor. A Crosshead Assembly is another component we would consider should be expected to last the lifetime of the compressor. If damage to a Crosshead does occur then it is usually as a result of some secondary failure, such as a Small End Bush seizure, as mentioned above. Crosshead Slippers are in most cases now, detachable to minimise the downtime required to change them, should the White Metal become damaged. These can then be repaired and re-used. With Crossheads with integral rubbing surfaces, the complete assembly would need to be removed for repair, usually requiring a complete assembly being kept as a spare. 3.3(A).3 Piston and Piston Rod Assembly The Piston Rod has probably suffered from more failures in the past, than any other of the components that would not be considered as “Wearing Parts”. Although the surfaces that enter the Packings do suffer from wear, they can be easily repaired by a variety of methods. The rest of the Rod should be designed to last the lifetime of the compressor.
Unfortunately, Rods have suffered failures for a variety of reasons which include: • Designed for unrealistically high Rod Loads (i.e. insufficient safety margin) • Locking devices coming loose or insufficient pre-tensioning resulting in the thread roots being subjected to unacceptable fatigue stresses. • Bending stresses induced by excessively high cylinder movement. • Incorrect material selection resulting in pitting or stress corrosion. • Shock loading due to liquid build-up in the cylinder. • Gross misalignment resulting in bending fatigue stresses. • Spare Piston Rods purchased from non-OEM sources with inferior properties. • Fatigue stresses resulting from excessively high Rod Loads due to operation on “OffDesign” conditions. • Corrosion fatigue as a result of operating on “Off-Design” processes. All of the above are preventable, some solutions being the responsibility of the OEM and some that of the End-user. Pistons very rarely suffer complete failures. Any that do occur are generally restricted to large Pistons that are either cast or fabricated. Typically, they will be due to defects such as porosity or post-weld cracking that has not been detected during the manufacturing stage. Much more common, is damage to the Piston caused by foreign matter entering the cylinder, such as broken valve plates and springs, or particles entrained in the gas, from upstream pipework and vessels. Usually this damage is superficial or easily repaired. The vast majority of Pistons last the lifetime of the compressor, so from the point of view of A & R, they should really be discounted. 3.3(A).4 Cylinder Assembly The following components are designed to last the lifetime of the compressor and as such should under normal circumstances have no effect on A & R. • • • • • • • • Main Cylinder Body Cylinder Mounting Piece (if fitted) Cylinder Head Valve Cages and Covers Cylinder Liner Fixed and Variable Clearance Pockets (if fitted) Coolant Space Covers All Studs, Bolts, Nuts and Fixings
In some arduous services the Liners can and do wear and need replacing, however, in the vast majority of cases they are never replaced. Similarly, with Studs, Bolts, Nuts and Fixings, O Ring Seals and Gaskets, these can suffer damage during maintenance, or corrosion in outdoor installations, however, the time involved with replacing these items should be minimal.
3.3(B) Major Component Design 3.3(B).1 Cylinder Valves These are without doubt the most important components in the whole compressor, from the point of view of A & R. If the Cylinder is considered to be the heart of the compressor, then Valves can be likened to the heart valves. The difference between them, however, is that Cylinder Valves have to operate more frequently and there is probably more that can go wrong with them! If we consider the classic Metallic Plate Valve, the following are just some of the problems that might occur: a).
Those that are within the Valve manufacture’s control. • • • • • • • • Incorrect Valve size selection (too small), resulting in high gas velocities leading to various types of failure. Incorrect calculations of the Suction Valve Seat thickness, resulting in fatigue failure. Incorrect Valve Lift selection, resulting in high impact loads leading to Valve Plate failure. Incorrect Valve Plate material selection, resulting in several different types of failure. Incorrect Closing and/or Damper Spring material selection, again resulting in several different types of failure. Incorrect selection of Closing and/or Damper Spring design resulting in Spring and/or Plate failures. Incorrect Heat Treatment of Valve Plate, Valve Seat or Spring material (when required), resulting in stress corrosion. Stress raisers left by the manufacturing process of Valve Plates or Valve Seats resulting in fatigue failures.
All these indicate what an extremely difficult task it is for the Valve designer to come up with the optimum design, considering the physical and commercial constraints. b).
Those that are outside the Valve manufacture’s control. • • • • • • • • The Valve being used on a gas for which it was never designed, resulting in various types of failure. The Valve being used for conditions for which it was never designed, resulting in various types of failures. The Valve Cage pre-load being lost, allowing the Valve to bounce, resulting in fatigue failure. Excessive pulsations in the Valve Chamber causing high impact loads and multiple impacts leading to Valve Plate failures. Excessive lubrication causing the Valve Plate to stick during opening and closing, again resulting in high impact loads leading to failure. Liquid carry-over, causing various problems which usually result in failure when the liquid is in “slug” form rather than vapour. Particulate matter, entrained in the gas, passing through the Valve, which at best will result in damage and reduction in performance. At worst could result in complete failure of the Plates and Springs. Incorrect tightening of the central Clamping Stud, causing it to loosen and the whole Valve assembly to vibrate, resulting in fatigue failure of some of the components.
Most of the above problems can and should be easily eliminated as factors that have an effect on A & R. In an ideal world and given no physical or financial constraints, all the other problems could be eliminated. Unfortunately, we live in a real world, with both physical and financial constraints. The Valve manufacture in conjunction with the OEM, must come up with a Valve design that will: a).
b).
c).
d).
e).
Be the smallest practical size Give optimum performance Have the lowest possible Valve Losses Be made from the cheapest possible materials. Last for a minimum of 3 Years running time.
Oh and of course, it would need to be competitive against those made by P I Rate Ltd., with his workshop at the bottom of his allotment, who makes damn good Valves, that are “Fit for Purpose”! Joking aside, I believe that the Cylinder Valve probably has as much effect on A & R, as all the other factors put together. In other words, if the Valves are 100 % Reliable, then you have eliminated 50 % of the factors affecting A & R. 3.3(B).2 Piston and Bearer (Rider) Rings As mentioned previously (3.1.4 ), there are two distinct types of Rings, those that are designed for Lubricated cylinders and those that are designed for Non-lubricated. In both cases I will assume that Non-metallic rings are used. If we assume that the Bearer Rings wear out before the Piston Rings (which is invariably the case) and hence are the components that will affect the Reliability, then there are fundamental differences between the way these have to work and the wear mechanism in Lube compared to Non – Lube service. With Lubricated service the Bearer Rings ride over a film of oil in the cylinder and provided this film is maintained then no contact should take place between the Rings and the Cylinder Liner surface. Theoretically therefore, assuming no deleterious materials are present in the gas, no wear should take place. Indeed, in the days before non-metallic rings were available, Bronze or White Metal pads were attached to the pistons. With the exception of some of the more arduous duties, these pads lasted for years and in some cases for the life of the compressor. Since the introduction of the Filled PTFE ring, specifically for Non-Lubricated service, these have also all but replaced the metal pads as a method of supporting the piston and rod weight, in Lubricated service. This has mainly been due to the aggressive marketing and selling efforts of the Non-metallic Ring suppliers. Nowadays, if say 3 years (26,280 Hours) life is achieved, it is considered very good, however, compared to what could be achieved, this is nothing special. It is now commonplace to offer sets of Piston and Bearer rings for Commissioning, 2 Year and Long Term Spares for Lubricated service. Although this is good for business for the Ring Vendors, it does nothing to enhance the reputation of the Recip as a Reliable Compressor.
In fairness, because in many cases now process gas compressors have to be designed for multi-service, many have to be designed for Non-Lubricated service, but will run for the majority of their life on Lubricated service. On Non-Lubricated service, where there is no choice but to use Non-metallic Rings, significant advances have been made in extending the life expectancy of a set of rings. It is generally accepted within the industry now, that anything less than 1 Year (8760 Hours) life, is unacceptable. Depending on process constraints however, a set of Rings can be changed in a matter of hours, so the effect on A & R can be minimal. 3.3(B).3 Main Gas Packings As with Rings there are two different types, Lubricated and Non-Lubricated. Also, both these types can be either Cooled or Non-Cooled. Basically the same fundamental principals apply as with Piston and Bearer Rings, however, due to their complex design, Packings have several more features that need to be designed and manufactured correctly to ensure that A & R is not compromised. The following are some of the more common problems that can occur: • Incorrect selection of the number of Sealing Elements, (Either too many or too few), resulting in various problems. • Incorrect Wearing Element material selection, resulting in rapid wear, erosion or breakage. • Incorrect Garter Spring selection, resulting in rapid Wearing Element wear or Spring failure. • Incorrect Garter Spring material selection, resulting in failure. • Overheating due to insufficient cooling, or no cooling where cooling is required. • Leakage past the Sealing Elements, the lapped Joint Faces or the “O” Ring Seals, for a variety of reasons. Again most of the above problems can and should be eliminated as factors affecting A & R. There is no reason why the Lubricated Packing should not be expected to last for well over 1 Year in all but the most arduous duties and in many cases up to the magic 3 Years. With Non-lubricated Packings it is a different story. As with Piston and Bearer Rings, the designer’s aim is to guarantee a minimum operating life of 1 Year. The Sealing Elements have to effect a seal around a Piston Rod that is reciprocating in the axial direction and in the radial direction, with a combination of vertical and horizontal movement. In addition to that, in most Packings the Sealing Elements are the segmental design, with typically 3 radial or tangential joints. It is clear, therefore that these Packings are much more difficult to design than typical Rotary Seals, working under similar conditions. 3.4 Materials Generally speaking the standard and quality of materials has improved significantly over the years. Modern production, testing and inspection techniques have all played their part in ensuring a better end product. Furthermore, the acceptance criteria have become much more stringent and many repair procedures that were previously allowed, are no longer permitted. Probably the one exception to the above are castings. Although the end product is probably better than it was say 20 years ago, the improvements are not so
marked. Virtually “defect free” castings can be produced using computer simulations of flow and cooling to design moulds and so on, however, these techniques are greatly add to the cost. It is possible on some special castings, such as Low Temperature Steel, for the design and inspection costs to be more than the casting. In this extremely competitive business, these techniques cannot always be employed. Fortunately, with Forgings and Bar material, the “special” has now become the standard, so cost is not such an issue. 3.5 Instrumentation Instrumentation can be split into two main categories. a).
Basic Instrumentation and Control These are essentially for ensuring the compressor operates on the duties for which it was designed. It allows operators to make minor adjustments, to the process, utilities, etc., to ensure this happens. It also helps ensuring that the correct action is taken, if any parameters deviate from design, to an extent that damage may result. b).
Condition Monitoring The main function of this can be split again into two distinct categories. 1).
Predicting potential failures, allowing action to be taken preventing them occurring, or minimising consequential damage. 2).
Monitoring the condition of wearing components, allowing accurate planning and scheduling of Downtime for replacement. The following parameters are commonly monitored: • • • • • • • • Suction and Discharge Valve Temperature Main Bearing Temperature Motor Bearing Temperature Motor Windings Temperature Main Gas Packing Temperature Crankcase Vibration Piston Rod Drop Main Motor Amps
The following parameters can also be monitored, but currently much less commonly. Some are monitored on a periodic basis and some only on experimental machines: • • • • • • • Cylinder Pressure / Piston Displacement Crankpin ( Big End ) Bearing Temperature Gudgeon Pin Temperature Cylinder Liner Temperature Cylinder Vibration Cylinder Valve Vibration ( Using ultrasonics ) Gas Density ( Molecular Weight )
• Piston Rod Runout ( Vertical and Horizontal ) Monitoring can be carried out local to the compressor, with integration into the customer’s D.C.S. system, or remotely in the main control room. Where there is a Service Contract in place, where very short call-out times are required, the compressor can be monitored remotely from the vendor’s facility, via modem and even satellite link. 3.6 Quality This is one of the most, if not the most important factors that can affect A & R. Quite simply, if the quality of the design is inferior, then even if the manufacturing process is perfect, then the component is not going to be Reliable. Similarly, if the design is perfect but the manufacturing quality is inferior, then again the net result will be the same. The importance of having good quality procedures in place and of course working to them cannot be stressed enough. Although, important to the Recip manufacturer, it is even more important to the Component manufactures. One unreliable, critical component can turn an inherently excellent machine, into a total disaster. The introduction of the ISO 9000 QA system has greatly improved matters and registration is almost a must now in the industry. 3.7 Pulsation and Vibration Control This is probably the most important subject when considering the A & R of the Reciprocating compressor. The consequences of lack of control can vary from minor problems with Cylinder Valves, to disastrous failures of the Compressor and associated Piping System. It is however, an extremely complex subject and unfortunately cannot be included in this paper. There are of course many other papers on the subject that have been published over the years. Suffice for me to say that with the huge array of knowledge and computer software currently available, there is no reason why Pulsations should have any adverse affect on A & R. 4 Factors Affecting A. & R. [ Outside the OEM’s Control ] The following are some of the factors that can have an affect, in some cases a major affect, on A & R, that are totally out of the control of the OEM. I have not discussed them in detail but I believe they are worth mentioning because with some of these, the blame for problems resulting in reduced A & R, can still be placed on the Recip. Operating the compressor on “Off-Design” Process Gases. Operating the compressor on “Off-Design” Process Conditions. Carry-over of liquid and/or particulate matter from the upstream process. Poor Maintenance and low inventory of Spares. Replacing components with substandard, Non-OEM Spares. Poor Piping design giving rise to unacceptable nozzle loads, resulting in misalignment and possible nozzle failures. • Acoustic and/or Mechanical Resonance in piping systems giving rise to excessive vibrations, resulting in a variety of problems. • Poor Foundation design, resulting in excessive movement of the compressor, piping vibration and deterioration of the foundations. • Specification and use of so-called “Fit-for-Purpose” designs, that turn out to be totally Unfit-for-Purpose. • • • • • •
5 Peter Brotherhood Limited’s Experience and Philosophy on Availability & Reliability 5.1 Eliminating Factors Affecting A & R As a relatively small company, we cannot compete on price alone so we need an edge over our competitors. Being perceived as a company whose products are very reliable, can give us that edge, so A & R is extremely important to us. In the late 1960’s there was a belief by some in our company, that our range of Horizontal Recip Compressors were “Over-Engineered” and as such too expensive. (A comment I might add, that I have heard on more than one occasion in more recent years, usually from non-engineers).
In particular, it was thought that the Frame itself was built like the proverbial “Brick-Built Outhouse”. We therefore embarked on a programme of redesign of the main components, so that they were “just strong enough” to withstand the loads for which they had been designed. The result proved to be a very costly exercise, both from the financial point of view and our reputation. Where machines were sold with Gas Loads at or approaching the Frame Rating, numerous problems resulted. Quite simply they were not capable of withstanding these loads. In my early days as a junior designer in the Compressor Department, I spent most of my time visiting sites where we had problems of one kind or another which could be attributed to excessive movement of the Compressor. In one embarrassing case, which I remember well, that Crankcase was being held together with two large “I” beams and a series of large tie bolts. Needless to say, this did not do too much for my moral or my confidence in the product, for which I had chosen to concentrate my design career. At the earliest opportunity, I decided to try to eliminate as many factors as I could, that were adversely affecting the A & R of our compressors. 5.1.1&2 Frames and Motion Work My initial aim was to try to eliminate, or at least reduce to an absolute minimum, the chances of any problems with the Frame and Motion Work, effectively removing them from the equation. Once again we embarked on a programme of a complete re-design our range of Recips, by calculation, Strain Gauge Testing and later, verification using FEA techniques. We also benefited from the knowledge of what had worked in the past and more importantly, what had not. After deciding on the Frame Ratings (Gas & Rod Loads) for each Frame, an Overall Factor of Safety was used each calculation, appropriate to the type of loading the component would be subjected. The same criteria were used for each Frame, to eliminate the anomalies that had existed between Frames. The result has been extremely successful, in that the initial aims have been met. Although we do still experience problems with the Frame and Motion Work from time to time, these are nearly always due to factors outside our control. Without a doubt, the key factor in this success was the selection of sensible Factors of Safety. Some may still argue that there may some components where the safety margin is too conservative, however, I would say it is better to be safe than sorry. Several other features
have been added which have further reduced the chances of problems with the Frame and Motion Work. All Crankshafts have either integral or bolt-on Balance Weights as standard, to give the optimum balance. On compressors with only 1 Cylinder, Balance Lines are fitted as standard, which match the reciprocating weight of working line. On the larger Conn-Rods, the Bolts can be tightened hydraulically, as an option. As standard all other Bolts are tightened using extension gauges, for more accuracy than say using torque spanners. To try to eliminate the possibility of Small End Bush failures, we have taken two courses of action. Firstly, we have adopted our own standard for the minimum allowable “Rod Reversal”. This criterion is far more stringent than that given in API 618, 4th Edition. Secondly, we have developed a special Bush design, where low Rod Reversals cannot be avoided for process reasons. Considerable care is taken to ensure the alignment of the Conn-Rod Assembly, the Crosshead Assembly and the Conn-Rod to Crosshead Assembly are within design tolerances, virtually eliminating this as a factor. Crosshead Slippers have the White Metal sprayed on to the Shoes, ensuring consistent thickness and perfect bonding. Unless these are damaged, they should last the lifetime of the machine. On all but our smallest frame, the Piston Rod attachment to the Crosshead is a “Bolt” type arrangement, with a specially designed Hydraulic Nut as standard. This has all but eliminated the chance of incorrect tightening or operator errors. It also allows for much quicker tightening and release. 5.1.3 Piston and Piston Rod Assembly Piston Rod materials are selected to suit the process conditions for every cylinder we produce. In some cases, this means different materials on the same compressor; however, it ensures that the optimum life is achieved. All Piston Rods are hard coated with Tungsten Carbide, using a H.V.O.F. process, as standard. Other coatings are selected for specific duties, where the standard may not be suitable. In many cases these days, Aluminium Pistons are used. In all but a few rare cases, these are machined from solid forgings, which mean we can virtually guarantee their integrity, unlike those made from castings. 5.1.4 Cylinder Assembly One thing that we do which may be unique in the industry, is that every cylinder we produce is individually designed, down to the last detail. The Cylinder diameter is calculated to the nearest 5 mm, or in the case of circulators (very low compression ratios), to the nearest 1mm. Most cylinders below about 400mm diameter are manufactured from solid steel forgings as standard, regardless of design pressure. Again, we can virtually guarantee the integrity of these forgings and have eliminated all the problems associated with small cylinder castings. (Some cylinders are still made from castings, in particular when special materials are required).
Clearance Volumes are calculated in each case and as the design layouts are produced on 3-D CAD workstations, these can be verified to within 1mm3.
5.2 Increasing A & R 5.2.1 Cylinder Valves The number, size and type of Cylinder Gas Valves are selected to give the optimum performance and operating life, for each specific duty. Obviously, we rely very heavily on the expertise of the Valve manufactures, such as Hoerbiger, as we no longer design and manufacture our own. They of course, have the benefit of the combined experience of most of the compressor manufactures. We would expect a minimum life expectancy of 1 Year and would consider anything less as unacceptable. 5.2.2 Piston and Bearer (Rider) Rings The design and selection of material for Piston and Bearer Rings, particularly for NonLubricated service, is a very important factor in determining the A & R of the compressor. Again, we try to be as selective as we can, whilst considering other factors, such as remaining competitive. Some of the more exotic materials now available, although offering extended life expectancies than the more commonly used materials; they are also considerably more expensive. On Lubricated service, we would expect a minimum life of 2 Years and 1 Year on Non-Lubricated service. 5.2.3 Main Gas Packings As with Valves and Rings, we select these for each individual cylinder. Again we rely heavily on the manufactures in choosing the optimum Packing, however, we would not consider any option that would give us any less than 1 Year life. In most cases we use Packings that are the cartridge type, where the complete Packing, except the end flange, can be removed as an assembly. This allows the Packing to be worked on in the workshop and reduces the actual Downtime to a minimum. 5.2.4 Materials Generally my comments regarding Materials have been made in the previous sections, however, I can say that we are continually looking to use the best available and regularly reviewing our supplier list, to ensure standards are maintained. 5.2.5 Instrumentation In most cases the scope and level of sophistication of Instrumentation is specified by the customer. However, we do add to this scope in our bids, where we believe there is a significant benefit, without a major cost impact. For example, when a single non-lubricated compressor is specified without any Piston Rod Drop measurement, we would offer this as standard. In some cases on critical service, we may recommend some level of instrument redundancy. 5.2.5 Quality We are registered to ISO 9001 by L.R.Q.A. and can CE mark our compressors, for export to the EU countries, as a guarantee of their compliance with the appropriate EU directives. We are continually looking at ways of improving the Quality of our products, either by the design or the manufacturing process. We try to use only subsuppliers of critical equipment, that are also ISO 9000 registered, to
ensure the overall Quality of our products is not compromised. 6 Peter Brotherhood Limited’s Philosophy on Maintainability Within our company quality procedure relating to A.R.M., we have included various statements which define the principals we try to adhere to when designing one of our packages, to ensure that Maintainability does not compromise Availability. Particular attention is paid to access routes and areas for maintenance. Once these have been defined, we make every effort to ensure these are not encroached by any equipment or auxiliary services, such as cooling water pipework, instrument impulse lines, electric cables and cable trays. The use of 3-D CAD assists greatly in this process, as the computer model can be viewed from all angles and the designer can “walk through” it, ensuring our criteria are met. The customer can also review the layout throughout the contract, to ensure it meets with their requirements. We try to adopt as many features in our packages that will keep the Downtime to a minimum. For example, the use of “Cartridge” type Packings, where the complete assembly is removed and replaced with a spare, so maintenance can take place in the workshop, whilst the compressor continues to run. The use of Hydraulic tightening, allows fast and accurate removal and refitting of components. 7 Conclusion I believe the key to improving the image of Recips is to eliminate as many factors from the equation relating to Availability and Reliability. This might seem a very obvious statement to make, however, if we compare the Recip to say a car engine, you would not expect to have to replace, for example, a Conn-Rod or a Cylinder Head during the lifetime of the engine. The same should be true with the Recip. Whilst End Users still have to replace major components on their Recips, it’s reputation as being unreliable will never change. With careful consideration, the factors which have an affect on A & R, could be reduced to a small number of truly “Wearing” components. If the time period between changeovers of these components can be extended, then obviously Availability improves. The mathematics are quite simple. If we take the case where a Scheduled Downtime is say 10 Days. If this is carried out every year, then the Availability is 97.3 %. However, if the period between can be extended to 3 years, then the Availability improves to 99.1 % ! Should 2 Recips be required, when only 1 Screw or Centrifugal compressor is specified? A single Recip with suitable instrumentation and monitoring systems should easily achieve a figure for Availability of 97 % or better, compared to about 99 % for a Screw or Centrifugal (Note: This is what I would expect, not necessarily what would be offered ).Surely, when considering total operating and maintenance (Lifecycle) costs, the Recip should still come out on top? The process of improving the image needs to involve the End Users. As mentioned earlier in this paper, there are many factors that can affect A & R that are outside the control of the OEM, but can still have an adverse effect on its reputation. As my brother mentioned to me recently, “Perception is everything these days. Very few decisions in this world today are based on fundamentals”. If we can ensure the Recip is perceived as Reliable, then maybe half the battle is won.
